Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Auto Closed
-
None
-
None
-
None
Description
inspired by QPID-2992
We need some documentation to describe the kinds of 'race' problems you can get into when creating and shutting down messaging systems involving clusters, store, and federation. There are 'race conditions' that can give the appearance of bugs that depend on, for example, the speed with which you shut down brokers and the order in which you shut down brokers in a multi-broker messaging system.
We should have a single place in the docs where we collect what we know about the most common of these problems, and best practices for avoiding them.